Picking P-wave arrival times from microseismic signals in tunnels with low signal-to-noise ratios without noise filtering

Microseismic (MS) signals detected in tunnels feature low signal-to-noise ratios (SNRs) due to the complexity of the construction environment. As a result, MS events of low energy are hidden in the background noise so that separating such events from the noise using filtering is impractical. In this work, a new method is proposed to identify the arrival times of P-waves in such circumstances that does not require noise filtering. We call it the instantaneous phase difference intensity of P-waves (IPDI_P) method. The key requirement of the method (which is naturally met when monitoring MS signals in tunnels) is that the source wave travelling path difference between the target and auxiliary signals is not larger than the wavelength of the P-wave. IPDI_P functions are constructed that correspond to specific characteristics between the target and auxiliary signals. We take the first peak equal to, or close to, the maximum value of the IPDI_P curve to correspond to the P-wave arrival time of the target signal. The amplitude distributions of the two main noises in a tunnel (electrical noise and mechanical vibration) are first analyzed in the time- and frequency-domains and then accurately simulated. Then, numerical simulations and a case study are made. The results indicate that the IPDI_P method is reliable in tunnels without noise filtering and its performance is better than that of more traditional methods (STA/LTA and AIC), especially for MS signals with SNRs that are less than 30 dB (so the P-wave's SNR is less than 16-24 dB). The error in the Pwave arrival time caused by mechanical noise (vibration) is larger than that caused by electrical noise. The principles underlying auxiliary signal selection are also discussed. Our study is an important step towards the development of a fully-automatic MS-based rockburst detection system for use in deep tunnels.